Pickett and David BMC Res Notes (2018) 11:231 https://doi.org/10.1186/s13104-018-3328-3 BMC Research Notes RESEARCH NOTE Open Access Global connectivity patterns of the notoriously invasive mussel, Mytilus galloprovincialis Lmk using archived CO1 sequence data Thomas Pickett and Andrew A. David* Abstract Objective: The invasive mussel, Mytilus galloprovincialis has established invasive populations across the globe and in some regions, have completely displaced native mussels through competitive exclusion. The objective of this study was to elucidate global connectivity patterns of M. galloprovincialis strictly using archived cytochrome c oxidase 1 sequence data obtained from public databases. Through exhaustive mining and the development of a systematic workfow, we compiled the most comprehensive global CO1 dataset for M. galloprovincialis thus far, consisting of 209 sequences representing 14 populations. Haplotype networks were constructed and genetic diferentiation was assessed using pairwise analysis of molecular variance. Results: There was signifcant genetic structuring across populations with signifcant geographic patterning of haplotypes. In particular, South Korea, South China, Turkey and Australasia appear to be the most genetically isolated populations. However, we were unable to recover a northern and southern hemisphere grouping for M. galloprovin- cialis as was found in previous studies. These results suggest a complex dispersal pattern for M. galloprovincialis driven by several contributors including both natural and anthropogenic dispersal mechanisms along with the possibility of potential hybridization and ancient vicariance events. Keywords: Invasions, Population, Dispersal, Haplotype, Mytilidae Introduction costs continues to decline, public data banks that archive Quantifying dispersal in marine environments has been sequences are growing at an exponential rate [3]. Tese a long standing challenge due to the difculty in track- data banks play a central role in the life sciences because ing large numbers of microscopic larvae within oceanic they allow for the reproducibility of published research, basins [1]. As a consequence, indirect methods have been which recently, has been a contentious issue in the life developed, the most common of which is population sciences [4]. In invasion genetics and indeed, population genetics. In marine invasion ecology, population genetics genetics as a whole, such resources remain surprisingly is often employed to track the dispersal of invasive spe- underutilized [3], despite the fact that they possess a cies but the dynamic nature of marine invasions caused wealth of spatio-temporal sequence data generated from by changes in vector strength, transient dispersal barriers a variety of projects [5, 6]. In this study, we attempted and stochastic factors, poses a challenge [2]. One poten- to elucidate global connectivity patterns of the inva- tial alternative could be the use of archived sequence sive Mediterranean mussel, Mytilus galloprovincialis data which possess a temporal element. As sequencing Lamarck, 1819 by repurposing archived cytochrome c oxidase 1 (CO1) sequence data from public databanks. Mytilus galloprovincialis is a relatively small marine *Correspondence: [email protected] bivalve (5–8 cm) that is native to the Mediterranean but Department of Biology, Clarkson University, Potsdam, NY 13699, USA © The Author(s) 2018. This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creat​iveco​mmons​.org/licen​ses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creat​iveco​mmons​.org/ publi​cdoma​in/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Pickett and David BMC Res Notes (2018) 11:231 Page 2 of 7 has aggressively extended its range to the Americas, Asia, An automated search was preferentially chosen over a southern Africa and Australasia [7–9]. Te primary vec- manual search due to the speed of sequence acquisi- tor responsible for the spread of this species includes tion, and its highly discriminative nature (a specifc code shipping, more specifcally the transportation of plank- is unlikely to pull duplicates, or ambiguous sequences). tonic larvae in ballast water of commercial ships and Te mitochondrial DNA marker, cytochrome c oxidase 1 attachment of byssal threads to ship hulls [10]. In addi- (CO1) was chosen due to its overrepresentation in pop- tion, the ease of culturing M. galloprovincialis along ulation genetic studies for this species relative to other with its palatability have resulted in the transplantation markers. Te following qualifers were incorporated into of stock populations for aquaculture purposes in difer- the coding script: ‘Mytilus’, ‘galloprovincialis’ ‘mitochon- ent regions of the world [10]. Te success of the species drial’, ‘CO1’. After scanning 2480 mitochondrial gene in its introduced range is due to inherent biological char- sequences, 322 CO1 fragments were recovered (date of acteristics that make it an aggressive invader, including original search: April 2016). For verifcation, each data- high fecundity and recruitment rates [11], broad thermal base entry was manually checked and discarded if: (i) tolerance [12] and resistance to desiccation and para- the sequence could not be linked to published research sites [13]. Te objective of this study was to assess levels (peer reviewed articles, technical papers or conference of genetic diferentiation of M. galloprovincialis across abstracts) and or (ii) the sequence could not be traced multiple global localities. We hypothesized that M. gallo- to a specifc geographic locality. In a few cases the min- provincialis would exhibit marked genetic diferentiation ing program recovered ‘CO1-like sequences’ which were between northern and southern hemisphere mussels but discarded. To avoid compiling duplicated sequences, overall very low levels of genetic diferentiation within database entries were cross referenced and any duplica- hemispheres due to repeated introductions. tions were also discarded. To confrm collection dates, sequences were cross-referenced to their corresponding Main text publications and in cases where no collection date was Materials and methods specifed, authors were contacted directly for confrma- Data mining and alignment tion. Based on these aforementioned flters, a fnally A workfow for repurposing repository sequence data tally of 209 sequences representing 14 distinct popula- was developed (Fig. 1). A mining program was frst coded tions were tagged as ‘useable’ for this study and they in C++ to search several DNA databases for M. gallo- were all accessible from the GenBank database (Addi- provincialis DNA sequences. Tese databases included tional fle 1: Table S1; Fig. 2a). A series of alignment algo- GenBank, the European Nucleotide Archive, DNA Data- rithms (CLUSTALW, MUSCLE and MAFFT) were tested base of Japan and the Barcode of Life Database (BoLD). on the compiled dataset set in Geneious ver 10.1.3 [14] Fig. 1 Workfow for CO1 sequence acquisition of Mytilus galloprovincialis from data mining to sequence alignment Pickett and David BMC Res Notes (2018) 11:231 Page 3 of 7 Pickett and David BMC Res Notes (2018) 11:231 Page 4 of 7 (See fgure on previous page.) Fig. 2 a Distribution map of Mytilus galloprovincialis CO1 sequences: 1—South Africa, 2—Northwest Pacifc (NP) China, 3—South China, 4—Greece, 5—Chile, 6—Portugal, 7—Spain, 8—Australia East, 9—Australia West, 10—New Zealand (Auckland Islands), 11—Tasmania, 12—Turkey, 13—British Columbia (Vancouver Island), 14—Korea (South). Map Credit: Reto Stöckli, NASA Earth Observatory. b Haplotype network for Mytilus galloprovincialis based on mtDNA—CO1 sequence data. Size of circles is representative of individuals with that haplotype. The smallest circles represent a haplotype frequency of one. Each connecting line between haplotypes represents one mutational step and perpendicular lines represent an additional mutational change. Dashed circles indicate distinct haplogroups and edited in BioEdit ver. 5 [15]. Te MAFFT algorithm Te ɸST value between northern and southern hemi- provided the highest quality dataset as measured by bp sphere localities showed low but signifcant genetic par- length. It was also chosen because to its incorporation titioning (ɸST = 0.11, P < 0.01) with 88% of the genetic of iterative refnement steps that corrects for accidental variation nested within individual hemispheres. Pairwise misalignments [16]. comparisons showed generally high ɸST values indicat- ing strong genetic diferentiation across all localities. Genetic diferentiation analyses Tasmania, South China and South Korea were the most To determine evolutionary relationships among haplo- genetically isolated populations, with all three generating types, a statistical parsimony network was constructed signifcant ɸST values of 0.80–0.93 when compared with using TCS ver.1.2.1 [17], with the fxed connection limit other populations (Table 1). Te least genetically difer- set to 95%. Genetic diferentiation across populations entiated populations were between the North Atlantic was calculated via pairwise ɸST comparisons that were and Mediterranean populations (ɸST = 0.07–0.26) and carried